docs: add gsoc'2025 project blog

[AkashKumar7902]

This pull request adds a new blog post to the project, detailing a Google Summer of Code (GSoC) 2025 project focused on building a semantic search engine for video content. The post provides a comprehensive overview of the technical approach, including the AI-powered ingestion pipeline, intelligent segmentation, LLM enrichment, and the hybrid search application.

New blog post addition:

• Added _posts/2025/2025-10-07-gsoc-semantic-video-search.md.md containing a detailed write-up of the GSoC project "Building a Semantic Search Engine for Any Video," including problem statement, technical solution, and outcomes.

Technical content highlights:

• Describes the multi-stage ingestion pipeline utilizing AI models for shot detection (TransNetV2), transcription and diarization (WhisperX), visual captioning (BLIP), and action/audio recognition (VideoMAE, AST).
• Explains the "Boundary Scoring" algorithm for segmenting video into meaningful scenes, and the use of LLMs for generating segment titles, summaries, and keywords.
• Details the hybrid embedding strategy for semantic search using ChromaDB, combining text and visual embeddings for improved relevance.
• Outlines the backend (FastAPI) and frontend (Streamlit) implementation for real-time semantic video search,